Hi,

Additionally, having the node root volumes on dedicated aggregates allows you to perform non-disruptive controller upgrades (headswaps) in cDOT using the aggregate relocation feature for the data aggregates.

Aggregate relocation, I assume, won't even be possible if the aggregate contains a node's root volume. I'm not even sure if basic takeover/giveback will work properly as it also makes use of aggregate relocation.

Best regards,
Filip

On Fri, Jun 5, 2015 at 11:02 PM, Brandon Kitchen <bkitchen@datalink.com> wrote:

Yes, there are some good reasons to have a dedicated root aggr even on older 7-mode systems.

For example, if you ever have to run wafliron on a really large aggr0, you could be down for quite a while.

Doesn't happen that often but it does happen. Better to be safe than sorry IMO.

On smaller configs (like 22x0 or 25x0 with 12 or 24 disks), it's not really feasible to dedicate disks to aggr0. That's about the only exception I'll make on this practice personally. 8.3 w/ ADP helps with this problem quite a bit (lack of efficiency in smaller configs).

--
Brandon Kitchen
Datalink
919-606-6766
www.datalink.com 
Datalink OneCall Support Phone: 800.291.3230
Follow us on: Twitter l Blog l Facebook





On Jun 5, 2015, at 4:44 PM, Fletcher Cocquyt <fcocquyt@stanford.edu> wrote:

I’m curious how many admins dedicate 3 disks for ONTAP aggr0/vol0 vs having one big aggr0 and vol0 residing on the same aggr as the other data volumes?

Is there a compelling reason to dedicate disks to ONTAP besides “best practice” ?

thanks







_______________________________________________
Toasters mailing list
Toasters@teaparty.net
http://www.teaparty.net/mailman/listinfo/toasters


_______________________________________________
Toasters mailing list
Toasters@teaparty.net
http://www.teaparty.net/mailman/listinfo/toasters